Opendoor Technologies (OPEN -13.59%) dazzled investors over the last three months like few other stocks. The online home-flipper jumped an incredible 1,400% over the last three months, going from a little over $0.50 a share to more than $10 at one point.

The rally began with hedge-fund manager Eric Jackson making the case that the stock could be the next Carvana, which jumped to almost 100 times its original price after nearly going bankrupt in 2022. That argument gained steam online and helped turn Opendoor into a meme stock, as it initially surged on high volume and no news.

Since then, the stock gained on real news. That includes the prospect of the Federal Reserve lowering interest rates next week and later in the year, and the company’s board overhauling its management team. In August, embattled CEO Carrie Wheeler stepped down; after hours on Wednesday, Opendoor named Shopify chief operating officer Kaz Nejatian as its new CEO, which sent the stock up 80% on Thursday.

Additionally, the company said that co-founders Keith Rabois and Eric Wu were rejoining the board of directors, and ventures associated with them were investing $40 million into Opendoor. It’s easy to see how that news would inject enthusiasm into the stock, especially after it was on the verge of being delisted by the Nasdaq stock exchange earlier.

However, nothing’s really changed for Opendoor as a business in the last three months. The company never reported a full-year profit, and the business is expected to shrink this quarter due to the weak housing market.

It’s still a high risk with a questionable business model. If you’re looking for a similar stock that can capitalize on falling interest rates, I think that Upstart Holdings (UPST 1.54%) is a better bet, and that it can outperform Opendoor over the next three years.

Upstart’s opportunity

Upstart has a number of things in common with Opendoor. Both went public around the same time in 2020, and initially surged out of the gate before plunging in 2022 as interest rates rose and tech stocks crashed.

Upstart is a loan originator. It uses artificial intelligence (AI) technology to screen applicants, producing results it claims are significantly better than traditional FICO scores. Once it creates a loan, it typically sells it to one of its funding partners, so it doesn’t keep the debt on its books.

Like Opendoor’s, Upstart’s business was struggling back in 2022, but the company revamped its business with the help of an improved AI model that increased conversion rates for its loans. Even in a high-interest-rate environment, it’s delivering strong revenue growth. And it’s now profitable based on generally accepted accounting principles (GAAP).

Revenue in the second quarter jumped 102% to $257 million, on a 159% increase in transaction volume. The company reported GAAP net income of $5.6 million, and for the full year, it expects that to be $35 million.

Upstart built its business around consumer loans, but it’s been expanding rapidly into auto and home loans. The home loan market, where it could potentially compete with Opendoor, is massive. In the second quarter, Upstart’s home originations grew nearly 800% from the year-ago quarter to $68 million. That’s still a small fraction of its business, but there’s clearly more growth ahead in the home loan market for Upstart.

Upstart vs. Opendoor

Upstart and Opendoor have similar market caps following Opendoor’s surge. Upstart is valued at $6.1 billion as of Friday, while Opendoor’s market cap is $6.7 billion.

Both companies are also chasing massive addressable markets, and are likely to benefit from lower interest rates.

However, Upstart is the only one of the two that has proven it can grow in a challenging macro environment, and its business now looks set for consistent profitability. At Opendoor, meanwhile, there are real questions about whether home-flipping can scale up as a business model and deliver a consistent profit. Notably, both Zillow Group and Redfin (a subsidiary of Rocket Companies) bowed out of the iBuying competition, finding it too difficult and prone to large losses.

Given those differences, despite the fanfare over Opendoor, Upstart looks like the better bet today. Over the next three years, Upstart looks set to be the winner of the two.

The opportunity is real for Brussels, and the stakes are high, as the EU continues to trail the U.S. on virtually every cutting-edge technology—including artificial intelligence. A recent BCG Henderson Institute report shows that that stricter immigration rules and deep funding cuts for academic research in the U.S. raise the possibility that top AI researchers, a large share of whom are not U.S.-born, could look to take their talents elsewhere. Repatriating those top European academics is an important step for European policymakers, but to catch up, the EU must also be able to attract talent beyond the European diaspora, which is only a small fraction of the globally mobile AI talent base.

To remake itself into a tech talent magnet, Europe needs to build an academic ecosystem more closely integrated with its industries, a necessary step to provide the career pathways and information flows needed to turn academic discoveries and inventions into business value. The cost of this transformation will be considerable, as publicly discussed in, for instance, the Draghi report. Only then can the EU’s investments in academia help generate longstanding economic and geopolitical returns for the bloc.

The opportunity for Europe must not be overstated 

The EU recently announced a €500 million allocation over the next two years to help attract foreign researchers. Member states have also launched their own initiatives, including France’s €100 million commitment to its “Choose France for Science” platform to attract international researchers, and Spain’s €45 million pledge to help lure scientists “despised or undervalued by the Trump administration.” 

If these investments are made with the sole aim of repatriating European AI talent in the U.S., they risk falling short. The U.S. is home to roughly 60% of the top 2,000 AI researchers in the world, only one-fifth of whom are originally from continental Europe. Even an exodus of historical proportions would cover only half of the current gap between the EU and U.S. shares of the top AI researchers. 

At top GenAI labs, such as OpenAI and Anthropic, only a very small fraction of AI specialists (less than 1 percentage point of the 25% of workers who have completed their undergraduate degree outside of the U.S.) completed their bachelor’s degree in the EU. The future pipeline of AI talent is no different: In 2023, the top 10 contributing countries of foreign-born PhD recipients in computer science and mathematics to the U.S. accounted for 80% of the total. But not one of those countries is in continental Europe.

The U.S. AI research ecosystem is overwhelmingly supported by talent from Asia, not Europe: 85% of U.S.-based foreign nationals in technical AI jobs at leading American labs hail from China or India. So do 60% of all U.S. computer science and math Ph.D graduates in the U.S. Iran, Bangladesh and Taiwan account for most of the rest. If the EU is serious about becoming a vibrant hub for global AI research talent, it needs to look eastward.

But current (and prospective) AI researchers often don’t see Europe as a top destination. BCG’s Talent Tracker shows that Germany does best among European countries, ranking 5^th globally as a “dream destination” for highly skilled talent, followed by France (9^th), Spain (10^th), and the Netherlands (16^th). The EU is not just less attractive than the U.S. (2^nd), but also Canada (3^rd), the UK (4^th), and Australia (1^st), and roughly on par with the UAE (11^th). European countries are by no means the only nations committed to boosting their own talent bases.

Part of the challenge is the lack of large EU academic institutions with strong AI credentials compared to other regions. None of the top 50 AI institutions worldwide (as ranked by Google Scholar’s H5 journal impact index) are in the EU. A strong institutional base for leading AI labs is essential to create the work environment capable of attracting the best and brightest. 

The EU needs to invest in its universities to improve its standing, but it must also look beyond academia to improve its entire innovation ecosystem. Nearly a third of non-U.S. AI specialists go to the U.S. because of its extensive opportunities for career growth, including entrepreneurial endeavors, a BHI survey of top tech talent recruiters found.

The need for a concerted strategy across academia and industry

To get started, European countries must improve academic compensation in critical fields related to AI, and technology more broadly. In Europe, even when adjusting for purchasing power parity, salaries at the associate professor level are half of those paid at top U.S. institutions. Europe also needs to increase grant availability for research. Public research grants for computer science and informatics at leading American AI institutions are double those available in Europe. Europe may get a boost however, if the U.S. goes through with proposed cuts to the National Science Foundation’s budget.

It’s well known that incentives for innovation matter. In the 2000s, a few European countries reformed their academic patenting laws to follow the U.S. model, where American universities hold patent rights and share commercialization profits with professors. But the reforms were not well tailored to the European context and led to a significant decrease in academic patenting (between 17% and 50% depending on the country).  

Furthermore, only about a third of patented inventions from EU universities and research institutions ever get exploited, largely due to their weak integration into innovation clusters that drive commercialization. Even the best EU innovation clusters, once again, fall outside the top 10 globally, with the U.S. accounting for four spots, and China three. To change that, it’s essential for European policymakers to help build stronger bridges between academia and industry to ensure that foundational research effectively fuels economic value creation.

That includes strengthening the startup and innovation ecosystem around universities themselves. The ultimate aim of attracting top AI researchers is not to simply catch up, but to skip ahead and produce the next IP breakthrough, which will only rise in importance as more AI models become commoditized. Coming up with the next big thing, however, requires an investment environment capable of supporting ambitious bets on potential breakthroughs coming out of academia. Countries like Canada and the U.K. serve as cautionary tales of AI research hotspots that have often struggled to translate academic breakthroughs into commercial successes, a leap successfully undertaken by large U.S. tech companies.

Many of the usual items in the European reform menu will also bolster the AI talent and innovation ecosystem. As the 2024 Draghi report on the future of European competitiveness noted, the integration of EU capital markets is vital, as is the removal of internal trade barriers that hamper early-stage startups’ growth. Between 2019 and 2024, AI venture capital investment in the EU was just a tenth of that in the U.S. It is no wonder then that nearly a third of European “unicorns” founded between 2008 and 2021 relocated elsewhere—usually to the U.S. 

But crucially, the list of reforms must also include strong incentives for AI adoption. At present, EU companies lag their U.S. counterparts in generative AI adoption by between 45% and 70%. Closing that gap will simultaneously help fuel European demand for specialized AI talent and create the economic opportunities beyond academia that are critical to attracting the world’s best and brightest.

Overconfidence could set back the EU 

The EU is right to want to lure researchers into its academic institutions that have historically pushed the frontier of AI. This will require revamping the academic ecosystem and more systematically translating academic breakthroughs into long-term economic and strategic leadership. 

But it would be wrong for European policymakers to assume that the erosion of U.S. attractiveness will organically lead to a talent windfall, predicated on their belief that Europe is the inevitable “next best” option. That will only be true if the region acts decisively to build its own, integrated, AI ecosystem capable of attracting the brightest minds from China, India, and beyond. In the AI race, as on many other fronts, the EU bears the risk of being too confident in its belief that it is entrenched in third place. That kind of complacency could very well accelerate the EU’s descent into the minor leagues of global innovation.

***

Read other Fortune columns by François Candelon.

François Candelon is a partner at private equity firm Seven2 and the former global director of the BCG Henderson Institute. 

Etienne Cavin is a consultant at Boston Consulting Group and a former ambassador at the BCG Henderson Institute.

David Zuluaga Martínez is senior director at Boston Consulting Group’s Henderson Institute.

Some of the companies mentioned in this column are past or present clients of the authors’ employers.

Out of Fireworks and into the fire

However, my start with Cerebras’s hosted Qwen was not the same as what I experienced (for a lot more money) on Fireworks, another provider. Initially, Cerebras’s Qwen didn’t even work in my CLI. It also didn’t seem to work in Roo Code or any other tool I knew how to use. After taking a bug report, Cerebras told me it was my code. My same CLI that worked on Fireworks, for Claude, for GPT-4.1 and GPT-5, for o3, for Qwen hosted by Qwen/Alibaba was at fault, said Cerebras. To be fair, my log did include deceptive artifacts when Cerebras fragmented the stream, putting out stream parts as messages (which Cerebras still does on occasion). However, this has been generally their approach. Don’t fix their so-called OpenAI compatibility—blame and/or adapt the client. I took the challenge and adapted my CLI, but it was a lot of workarounds. This was a massive contrast with Fireworks. I had issues with Fireworks when it started and showed them my debug output; they immediately acknowledged the problem (occasionally it would spit out corrupt, native tool calls instead of OpenAI-style output) and fixed it overnight. Cerebras repeatedly claimed their infrastructure was working perfectly and requests were all successful—in direct contradiction to most commentary on their Discord.

Feeling like I had finally cracked the nut after three weeks of on-and-off testing and adapting, I grabbed a second Cerebras Code Max account when the window opened again. This was after discovering that for part of the time, Cerebras had charged me for a Max account but given me a Pro account. They fixed it and offered no compensation for the days my service was set to Pro, not Max, and it is difficult to prove because their analytics console is broken, in part because it provides measurements in local time, but the limits are in UTC.

Then I did the math. One Cerebras Code Max account is limited to 120 million tokens per day at a cost equivalent to four times that of a Cerebras Code Pro account. The Pro account is 24 million tokens per day. If you multiply that by four, you get 96 million tokens. However, the Pro account is limited to 300k tokens per minute, compared to 400k for the Max. Using Cerebras is a bit frustrating. For 10 to 20 seconds, it really flies, then you hit the cap on tokens per minute, and it throws 429 errors (too many requests) until the minute is up. If your coding tool is smart, it will just retry with an exponential back-off. If not, it will break the stream. So, had I bought four Pro accounts, I could have had 1,200,000 TPM in theory, a much better value than the Max account.