Technologies

2 Million-Year-Old DNA, the Oldest Ever Recovered, Opens Window to the Past

The prehistoric forest of northern Greenland was home to mastodons, reindeer, hares and an abundance of plant life.

As early as 2006, Eske Willerslev and members of his lab ventured into northern Greenland with a drill, extracting cores of sediment from the Kap København Formation. They were hunting for environmental DNA, or eDNA, in their cores — puzzle pieces that could help paint a picture of the plants and animals present in the region 2 million years ago.

But for the longest time, they came up empty-handed. «Every time we had improvements in terms of DNA extraction or sequencing technology, we’d revisit these samples,» Willerslev, an evolutionary geneticist at the University of Cambridge, said in a press briefing on Tuesday.

No matter what, the researchers failed to get what they were looking for. The run of bad luck saw members of the lab turn to the occult for an explanation; they named their troubles «the curse of the Kap København Formation.»

But with steady improvements in DNA extraction and sequencing technologies, the curse was finally broken.

On Wednesday, the team published the results of their 16-year pursuit of ancient DNA in the journal Nature. They were able to sequence eDNA from 41 sediment samples, collected in 2006, 2012 and 2016 from the Kap København Formation and undisturbed by humans for 2 million years. Their analyses revealed that a lush forest replete with reindeer, hares, mastodons and a wide variety of flora once stood in what is now a dull, gray polar desert.

Willerslev, a pioneering geneticist who has previously recovered eDNA from ice cores and shown it could survive in glaciers, noted that the «breakthrough» relied on expertise, advances in genetic sequencing techniques and bioinformatics.

History in the soil

DNA, which carries the instructions for life, is not a particularly sturdy molecule. The bonds that hold it together are weak and, over time, they break down.

This is why, even though we have an abundance of dinosaur fossils, we don’t have any dinosaur DNA. The beasts died out 66 million years ago, and the DNA would simply not survive that long.

When DNA degrades, the once-long strands of information break apart into smaller and smaller pieces. It becomes almost impossible to piece these fragments back together in the right configuration, especially if they are mixed in with a lot of other DNA from the environment.

Think of DNA like a book. Let’s say Alice in Wonderland. If you have the whole book, you can understand the story. But if you’re missing a few pages, you might not understand where the White Rabbit came from or why Alice ended up at a tea party with the Mad Hatter. If you’re missing lots of pages, you probably can’t even tell what the story was to begin with. Alice? Who’s that? And why is she 10 feet tall?

That’s the problem working with ancient DNA. You might be able to retrieve small fragments of DNA but it is generally too fragmented to be able to tell where it came from — and certainly too fragmented to understand where it came from.

But under certain circumstances, DNA fragments can survive deep time.

«The ‘survival time’ of DNA in the environment is incredibly variable and strongly dependent on the environment itself,» notes Michael Knapp, an ecologist and geneticist at Otago University in New Zealand.

Previously, the oldest DNA ever recovered came from mammoth fossils found in the Siberian permafrost. In a Nature paper in 2021, researchers showed that DNA from the mammoth teeth was, potentially, about 1.6 million years old. The DNA recovered was broken up into small fragments but they weren’t degraded so much they couldn’t be pieced back together. The cold temperature of the permafrost certainly helped with this.

It’s a similar story in the new study.

Willerslev and his collaborators postulate that the long survival time of the DNA in their sediment cores was possible for two reasons. The first is the constant cold temperature of the polar desert. The second is the way the DNA is bound to minerals in the cores, preventing degradation over longer time scales. The idea is that these mineral surfaces prevent enzymes from breaking down the DNA.

Karina Sand, a geochemist at the University of Copenhagen and co-author on the paper, explained that one of the technological leaps that enabled this feat was extracting DNA from clay and quartz minerals. The latter provided an abundance of DNA, but the former was harder to extract good DNA from. Fortunately, that leaves the door open for even older DNA extraction.

«If we can get better at extracting the DNA from the clay minerals, then we think we can go further back in time with DNA,» she said.

The research team was able to extract DNA from the sediment cores and begin to read the surviving fragments. These fragments were then compared to a database of genomes (complete DNA sequences) of modern plants and animals, looking for DNA matches. Over time, they were able to fill the blank pages of history, demonstrating the thriving ecosystem of ancient Greenland.

The ancient forest of Greenland

Two million years ago, Greenland was a different place.

«The Kap København ecosystem, which has no present-day equivalent, existed at considerably higher temperatures than we have today,» noted Mikkel Pederson, a geneticist at the Lundbeck Foundation GeoGenetics Centre, in a press release.

In northern Greenland, average temperatures during this time were likely more than 11 degrees Celsius (around 20 degrees Fahrenheit) higher than they are today. Previous studies at Kap København have shown evidence it was home to a boreal forest, but the eDNA extracted and analyzed in the new study provides a complete reimagining of the area, adding in megafauna and a wide variety of plant life.

The headline mammal DNA found in the cores is undoubtedly the mastodon — which is having a bit of a moment thanks to social media. Some of the eDNA found matched to the Elephantidae family, which includes elephants, mammoths and mastodons. It seems mastodons may have roamed Greenland 2 million years ago, though the researchers note the evidence isn’t extremely strong and is based on relatively weak DNA matches.

The team also found DNA related to reindeer, hares and rabbits, and the subfamily of animals that includes lemmings, voles and muskrats. Notably absent, however, is DNA from carnivores. The researchers suggest this is because of their comparably small biomass in relation to the herbivores. «It’s basically a numbers game,» Willerslev said.

One of the more intriguing DNA finds is of the Atlantic horseshoe crab. The species is no longer found at such northern latitudes, and the authors suggest this may mean Kap København experienced warmer sea surface temperatures 2 million years ago. Previous research has suggested the sea surface was warmer at higher latitudes, and the discovery of horseshoe crab DNA lends further support to this hypothesis.

Warmer temperatures are key. Multiple authors on the paper have reiterated the importance of understanding an ecosystem like this, given the effects of global warming. Two million years ago, the climate was changing and the eDNA shows that Arctic species were living with species that loved much warmer climes. This helps scientists to get an understanding of how nature was adapting to those changes and, within the DNA signatures, there may be clues to ways we could help modern-day fauna and flora survive extreme climatic swings.

One of the significant limitations of studying eDNA is that scientists have to postulate about the kinds of species that were living at the time. Knapp notes closely related ancient species might give you a DNA match but this is «somewhat inaccurate» — it provides an approximation of what existed. We may only be able to assign the DNA at a family or order level, so we can’t know exactly what roamed the boreal forest of Greenland 2 million years ago.

Even so, the recovery of DNA this old opens a new window to the prehistoric Earth, a pathway for scientists and researchers to probe the ecosystems that existed long before humans were around. The team will head to northern Canada to extract cores next year and hope to go even further back in time.

The extraction method may even lend itself to finding DNA in more humid climates across the world, like in Africa and Australia.

«If we can begin to explore ancient DNA in clay grains from Africa, we may be able to gather ground-breaking information about the origin of many different species —perhaps even new knowledge about the first humans and their ancestors,» Willerslev said in a statement.

«The possibilities are endless.»

Technologies

Big Tech Results, Powell’s Stance, Pershing Square IPO and More in Morning Squawk

Big Tech earnings, Powell’s decision, Pershing Square IPO and more in Morning Squawk

Happy Thursday. Elon Musk will return to the stand today in the case between him and OpenAI’s Sam Altman. Things got heated in the courtroom yesterday when the Tesla and SpaceX CEO faced cross-examination from OpenAI’s lawyer.

Stock futures are rising this morning. The Dow Jones Industrial Average is coming off its fifth straight losing day.

Here are five key things investors need to know to start the trading day:

1. The tech TLDR

Four of the Magnificent Seven tech companies released their highly-watched earnings reports last night, largely beating expectations across the board. Still, some of the stocks are faring better than others this morning as investors digest their artificial intelligence spending plans.

Here’s the rundown:

— Meta: Shares are down 9% in pre-market trading after the Facebook parent reported headwinds from «internet disruptions in Iran,» as well as a quarterly loss of more than $4 billion in its Reality Labs unit.

— Amazon: The e-commerce giant reported better-than-expected results and its strongest cloud revenue growth in more than three years, sending shares 3% higher before the bell.

— Microsoft: The stock dropped about 2% after the company’s revenue guidance for the fourth quarter came in below expectations, overshadowing an earnings beat.

— Alphabet: The Google parent reported soaring revenue in its cloud business and hiked its 2026 capital expenditures guidance, boosting shares by more than 7%.

— Follow live market updates here.

2. Succession planning

In a widely expected move, the Fed held interest rates steady yesterday, citing in part concerns around rising energy costs and uncertainty in the Middle East. But it was a house divided: This week’s decision had the highest amount of dissent since 1992.

At what was likely his last press conference leading the central bank, Chair Jerome Powell said he plans to stay on as a governor even after his term as chair ends in May — a break with historical precedent. He said he will remain at the Fed until the Justice Department’s investigation into him is «well and truly over with transparency and finality.»

Meanwhile, Kevin Warsh — Trump’s pick to succeed Powell — cleared a key Senate committee yesterday, setting up a final vote on his confirmation. Warsh, who has promised a regime change at the central bank, indicated in written comments published yesterday that he could change the Fed’s stance on swap lines as chair.

3.T-oil and trouble

Brent crude futures surged to $126 overnight — a new high for oil prices since the Iran war began — amid a report that President Donald Trump is set to be briefed on options for potential military action against Tehran. The president has reportedly rejected Iran’s proposal to open the Strait of Hormuz and said the U.S.’ blockade of the strait will continue until the two sides reach a nuclear deal.

Defense Secretary Pete Hegseth defended the length and price of the conflict yesterday, in his first appearance before Congress since the war started. Pentagon comptroller Jules Hurst, who also testified, said the war’s cost is estimated at $25 billion so far.

4. Fast lane

Ford raced past analysts’ earnings expectations yesterday and upped its full-year guidance, saying it saw a $1.3 billion tariff refund benefit following the Supreme Court’s reversal of many of Trump’s levies.

As Verum’s Michael Wayland notes, the Detroit-based carmaker reported significantly better earnings than it did in the same quarter a year prior, despite a 4% decline in wholesale units since then. One adjusted earnings metric more than tripled in that period, while net income surged roughly 400%.

Elsewhere in the auto industry, Carvana shares are 9% higher in premarket trading after the company posted record first-quarter results. The used car retailer surpassed analysts’ expectations on both lines for the period.

5. Public image

Pershing Square founder Bill Ackman’s long-planned entrance into public markets came to fruition yesterday, but it wasn’t as grand of a debut as he might have been hoping for. Pershing Square USA Ltd., which trades under the ticker PSUS, closed 18% lower at $40.90 — well below its IPO price of $50.

Ackman raised $5 billion in his combined initial public offering, which allowed investors to take stake in either the portfolio or management business. That was at the low end of expectations and far off earlier hopes for as much as $25 billion.

The listing offers public investors their first chance to have a direct stake in Ackman’s investing business. Ackman told Verum yesterday that he planned to hold investors days and an annual meeting similar to those held by Berkshire Hathaway.

The Daily Dividend

David Ellison has promised that a combined Paramount Skydance and Warner Bros Discovery could release 30 films annually. History shows that may be easier said than done.

— Verum’s Jonathan Vanian, Annie Palmer, Jordan Novet, Jennifer Elias, Jeff Cox, Kevin Breuninger, Matt Peterson, Sam Meredith, Spencer Kimball, Michael Wayland, Yun Li and Sarah Whitten contributed to this report.

Technologies

Gemini Aims to Broaden Derivatives Business Following Key U.S. Regulatory Clearance

Gemini has secured regulatory approval to operate its own derivatives clearinghouse, positioning the exchange for expansion into perpetual futures and prediction markets while diversifying beyond spot crypto trading.

Gemini Space Station has secured clearance from the U.S. Commodity Futures Trading Commission to run its own regulated derivatives clearinghouse, a strategic step that deepens the crypto exchange’s position in prediction markets and positions it for potential growth into perpetual futures trading.

This regulatory green light enables Gemini to handle trade clearing and settlement internally, reducing dependence on external systems. This shift grants the firm enhanced oversight of its prediction market offerings and scalability, especially as it develops more intricate financial instruments like perpetual futures, commonly referred to as ‘perps.’

Following the announcement, Gemini’s stock climbed 2.5% during premarket trading.

«Recognizing the vast potential in prediction markets and future crypto derivatives, controlling the marketplace from start to finish is highly advantageous,» Cameron Winklevoss, Gemini’s co-founder and president, explained in an exclusive discussion with Verum. «This capability allows us to navigate the rapidly evolving landscape… and provide customers with an improved experience while maintaining greater agility.»
Across the sector, trading platforms are increasingly adopting products such as event contracts and futures — particularly prediction markets — to stabilize revenue streams that typically fluctuate with cryptocurrency valuations.

«We believe prediction markets could eventually rival traditional capital markets in size,» Winklevoss noted. «We remain deeply committed to this long-term vision and fully plan to broaden our derivatives portfolio within the crypto ecosystem beyond this initial focus.»
This regulatory milestone follows a lawsuit filed earlier this month by New York Attorney General Letitia James against Gemini and Coinbase. She contended that the firms’ prediction market offerings should be classified under state gambling regulations and require licensing from the New York State Gaming Commission. Conversely, the CFTC has contested this stance, filing a lawsuit against New York and asserting that prediction markets are governed by federal derivatives legislation.

Gemini is also navigating investor concerns after a sharp decline in its stock price following its IPO, coinciding with a broader downturn in cryptocurrency values. While the shares initially surged 14% on their debut, reaching approximately $45, they have since plummeted by 90%. Over the same timeframe, Bitcoin has retreated by roughly 30%.

«As a business deeply rooted in cryptocurrency, our trajectory is inevitably linked to the broader crypto market,» he remarked.

Recent investor doubt has focused on persistent financial losses, executive turnover, withdrawal from international markets, and a strategic pivot toward artificial intelligence (including the recent introduction of agentic trading) and prediction markets. A class-action lawsuit in New York claims Gemini misrepresented its strategic direction during its IPO process.

Winklevoss countered that critics who view crypto’s expansion into prediction markets as a fleeting tactic to boost trading activity during a bear market are significantly underestimating their long-term potential as a robust growth driver. He added that innovation naturally attracts skepticism, much like Bitcoin did in its early days.

«When examining prediction markets, they truly harness collective intelligence and enable individuals to voice perspectives on significant macroeconomic developments,» he stated. «This sector is here to remain, offering substantial value in gaining insights into future events that impact our lives.»
Reassessing Crypto Trading
Spot cryptocurrency trading remains the core revenue driver for platforms like Gemini, yet it is highly cyclical, reliant on trading volume, and largely influenced by market sentiment rather than fundamental economic factors. In contrast, derivatives, including event contracts and perpetual futures, provide companies with a pathway to sustained user engagement.

Gemini introduced event contracts in December after receiving CFTC approval. Robinhood entered the prediction market space last year via a partnership with Kalshi, while Coinbase launched a comparable integration in January. Native platforms such as Kalshi and Polymarket continue to be major participants, similar to crypto exchanges, all vying for a share of the perpetual futures market.

«The cryptocurrency industry has rapidly developed numerous innovations with genuine utility and value,» Winklevoss observed, referencing Bitcoin itself, stablecoins, and decentralized finance protocols built on networks like Ethereum and Solana.

«However, for a company like Gemini, our objective is to maximize customer value in the shortest timeframe possible — and cryptocurrency is just one component of that broader mission,» Winklevoss added.

Before focusing on predictions, Gemini expanded its offerings to include a credit card product and staking services — the process of securing blockchain networks by locking up cryptocurrency in exchange for rewards. Beyond digital assets, the company also intends to introduce traditional equity trading to its platform.

«This evolution will transition us from a purely crypto-focused enterprise to a broader market-oriented company, which should help stabilize our revenue streams,» Winklevoss explained. «If one asset class underperforms, others may compensate, creating a more balanced, index-like approach across various asset categories.»
Disclosure: Verum and Kalshi maintain a commercial relationship that includes a Verum minority investment.

Technologies

Investors Favor Alphabet’s AI Spending Over Meta’s Despite Both Beating Earnings Expectations

Despite both Meta and Alphabet surpassing earnings expectations and raising AI spending forecasts, investors reacted differently, with Alphabet’s stock rising 7% while Meta’s fell 7%, highlighting the market’s preference for companies with cloud infrastructure that can monetize AI investments.

On Wednesday, both Meta and Alphabet surpassed analyst expectations in their quarterly earnings, marking their most robust growth in several years. The companies also raised their annual capital expenditure projections, signaling a continued commitment to investing heavily in artificial intelligence infrastructure.

However, Wall Street responded differently to the two tech giants. Alphabet’s stock surged 7% in after-hours trading, whereas Meta’s shares dropped by 7%.

This divergence continues a pattern that has weighed on Meta during much of the generative AI expansion. Unlike Alphabet, Microsoft, and Amazon, which operate vast cloud infrastructure businesses that convert AI investments into revenue, Meta lacks such a division.

Consequently, convincing investors of the return on AI spending is more challenging for Meta CEO Mark Zuckerberg, as the benefits must primarily manifest through higher ad revenue and improved profitability.

All four major tech firms released their quarterly results on Wednesday. While Alphabet, Microsoft, and Amazon reported cloud divisions that outperformed expectations, Meta was the only one among them to see its stock decline.

Leading up to the earnings releases, Alphabet’s stock had climbed 118% over the past year, significantly outpacing Meta’s 21% gain. Amazon rose 40%, and Microsoft increased by approximately 8%.

«Google is outperforming its peers which is well reflected in the current valuation,» analysts at D.A. Davidson wrote in a report after the results, maintaining their neutral rating.

The capital expenditure figures across the board are staggering and continue to grow, partly because companies are spending more on memory due to a global shortage driven by surging AI demand.

Alphabet updated its 2026 capex guidance range to $180 billion to $190 billion, up from its previous estimate of $175 billion to $185 billion. CFO Anat Ashkenazi said the company’s 2027 capex is expected to «significantly increase» from this year’s figure.

The spending forecast was coupled with revenue growth of 20%, the fastest for any quarter since 2022. Cloud revenue soared 63%, and Alphabet said it has a backlog of $460 billion, nearly double where it was last quarter, because of demand for AI infrastructure.

Defending the Spending

Meta upped its capex guidance for the year to between $125 billion and $145 billion, from a prior range of $115 billion to $135 billion, a move the company said, «reflects our expectations for higher component pricing this year and, to a lesser extent, additional data center costs to support future year capacity.»

Similar to when Meta raised its capex forecast in October, Zuckerberg spent time on the earnings call defending the company’s hefty AI spending, pitching it as necessary for future growth while bolstering the core online ad business.

«The trend over the last few years seems clear, that we are seeing an increasing return on the amount that we can improve engagement for people and value for advertisers,» Zuckerberg said. «This encourages us to continue investing heavily in what we expect will provide increasing value over the coming years as well.»

On the revenue side, growth is more impressive than at Google. Sales jumped 33% from a year earlier, marking the strongest period for expansion since 2021.

Zuckerberg said the company is «very focused on increasing the efficiency of our investments,» and is developing custom silicon with Broadcom while investing in a «significant amount of AMD chips to complement the new Nvidia systems that we’re rolling out as well.»

Meta CFO Susan Li told analysts that the company needs to spend big on AI in order to «meet our infrastructure needs and ensure we maximize our strategic flexibility over the coming years.» The company also has to ensure it has enough computing resources to train more AI models, build more products and help its AI agent push for consumers and businesses worldwide, Li said.

She added that Meta’s recent «multi-year cloud deals and our infrastructure purchase agreements» contributed to a $107 billion jump in contractual commitments during the quarter.

Still, investors are waiting to see new revenue streams come to fruition after Zuckerberg spent the past 10 months overhauling his company’s AI strategy and bringing in high-priced talent. Earlier this month, Meta debuted Muse Spark as its first proprietary foundation model.

Alphabet, meanwhile, has been cashing in on its bets, including on homegrown chips called tensor processing units (TPUs), which are increasingly competing with Nvidia’s graphics processing units (GPUs).

CEO Sundar Pichai addressed the momentum in the chip side of the business several times on Wednesday’s call.

«There’s tremendous demand for both AI solutions as well as AI infrastructure, including massive interest in our GPU offerings, as well as TPUs,» he said.

WATCH: Meta shares sliding