Google shut down AlphaFold, DeepMind's protein-folding AI tool, two weeks ago because it wasn't profitable, despite being one of artificial intelligence's most significant scientific achievements. The decision came after the project's leader, Demis Hassabis, insisted on making the tool's code and database freely available to the scientific community, limiting Google's ability to monetize the breakthrough.

AlphaFold represented perhaps the strongest evidence that AI could deliver on Big Tech's grand promises to cure cancer and solve humanity's greatest scientific challenges. Hassabis won the 2024 Nobel Prize in Chemistry partly for the tool's achievements. Yet Google, which is spending $200 billion in 2026 alone on new AI infrastructure, decided the project wasn't worth keeping alive.

What Made AlphaFold So Important?

Proteins are chains of molecules called amino acids that fold into specific three-dimensional shapes. These shapes determine how proteins function in the body. Misfolded proteins directly cause certain diseases and play a role in Alzheimer's, Parkinson's, and some cancers. For decades, scientists struggled to predict a protein's structure even when they knew its amino acid sequence.

AlphaFold solved this problem. The first version won a major protein-folding prediction competition in 2018. AlphaFold 2, released two years later, dominated the same competition so thoroughly that it helped earn Hassabis his Nobel Prize. The latest version, AlphaFold 3, could predict even more complex structures involving proteins, DNA, RNA, and other molecules.

The tool's capabilities were genuinely impressive. But they were also narrow. AlphaFold couldn't generalize beyond the specific tasks it was trained to perform. While successive versions became more powerful, that improvement came from human developers designing more sophisticated models with better training data, not from AlphaFold improving itself. The tool also couldn't explain why proteins fold the way they do, and it couldn't reliably predict the structure of a significant number of proteins.

Why Did Google Kill Its Most Celebrated AI Project?

According to reporting from the Financial Times, Google's senior executives grew frustrated with Hassabis because they saw him as too focused on scientific achievement and not focused enough on commercial returns. The tension centered on Hassabis's decision to make AlphaFold's code and database publicly available.

"While Hassabis is widely admired for his scientific leadership, several people familiar with Google's thinking said senior executives had become frustrated by what they saw as his lesser focus on the commercial demands of the company's AI business. His decision to make AlphaFold freely available became a source of tension inside Google, according to a person familiar with the matter. The project generated little commercial return despite the scale of its investment," the Financial Times reported.

Financial Times reporting cited in Source 1

A week after shuttering AlphaFold, Google removed Hassabis from his role directing DeepMind, giving him a face-saving "promotion" to a powerless position. The message was clear: scientific achievement matters less to Google than profit.

How This Contradicts Big Tech's AI Narrative

For years, tech executives have justified massive AI investments by promising transformative scientific breakthroughs. Alphabet President Ruth Porat said last fall that "we should be able to cure cancer in our lifetime" thanks to AI. Demis Hassabis himself suggested in a 2024 60 Minutes interview that "we can cure all disease with the help of AI".

Ruth Porat

These promises served a purpose. They helped convince policymakers and the public to support Big Tech's vision of an AI-driven future. They provided moral justification for the enormous resources being poured into AI development. But AlphaFold's shutdown reveals the gap between rhetoric and reality.

Google's decision exposes several uncomfortable truths about how Big Tech prioritizes AI development:

• Profit Over Science: Google killed its most celebrated AI achievement not because it failed scientifically, but because it couldn't generate commercial returns. The tool worked brilliantly at its intended task; it simply wasn't proprietary enough to make money.
• Narrow Applications Limit Value: AlphaFold's breakthrough was real but specialized. It solved protein folding but couldn't be easily adapted to other scientific problems. This narrow focus meant limited commercial applications and limited ability to recoup Google's massive AI investments.
• Open Science Threatens Monopolies: Hassabis's commitment to making AlphaFold freely available was the core issue. Had he agreed to keep the technology proprietary, Google would have enjoyed a commercial monopoly. Instead, the scientific community got access to the tool, and Google got little financial benefit.

What This Means for AI's Future in Medicine

Google's actions follow a familiar pattern in corporate America. Myriad Genetics famously filed patents on human genes it discovered, including two genes linked to breast cancer, to maintain a monopoly on genetic testing. The Supreme Court ended that practice in 2013 by ruling that human genes cannot be patented. But drug companies still patent cancer treatments and charge patients enormous prices.

The AlphaFold shutdown suggests that Big Tech will follow the same playbook. If AI tools can help cure disease, they will likely be kept proprietary and expensive rather than made freely available to the scientific community. The promise of AI curing cancer was always conditional on profitability. Now that condition has been made explicit.

How to Interpret Big Tech's AI Promises Going Forward

The AlphaFold story offers a roadmap for evaluating future claims from tech companies about AI's potential to solve humanity's greatest challenges:

• Follow the Business Model: Ask whether the company plans to keep the technology proprietary or make it open source. Proprietary tools are more likely to be developed and maintained, but they may be expensive or inaccessible to those who need them most.
• Distinguish Hype From Reality: AlphaFold was a genuine scientific breakthrough, but it was also narrow in scope. When tech executives promise that AI will "cure all disease," remember that even the most celebrated AI achievements have significant limitations and cannot easily generalize to new problems.
• Remember Profit Incentives: Big Tech companies are not charities. When they promise that AI will benefit humanity, ask yourself whether those benefits align with their financial interests. If they don't, the promises are likely to evaporate when profitability is threatened.

The irony is almost too perfect. Google, like the rest of the AI industry, spent years convincing the world that massive AI investments were justified by the promise of scientific breakthroughs. It then sidelined the scientist who actually delivered one of those breakthroughs because he dared to share it with the world. The charade, as one observer noted, can now end. Big Tech does not care about curing cancer unless it boosts profits.