Anthropic releases Opus 4.7 and Claude Design, Tesla completed its most important chip, Cursor valued at over $50 Billion

Each week, AI isn't just getting smarter, it's getting more capable, more embedded, and more consequential. This week, Anthropic shipped Opus 4.7, OpenAI made a direct bet on curing diseases faster, a three-year-old startup became worth $50 billion, Tesla completed a chip that signals it's no longer just a car company, and America's data centre buildout hit a wall it didn't see coming. Here's what happened, and what each story actually means.

Anthropic Just Released Its Most Powerful Public Model Ever, Claude Opus 4.7

On April 16, Anthropic released Claude Opus 4.7, its most capable model so far. The update brings a 13% improvement in coding benchmarks, resolves three times more real-world production tasks, supports high-resolution vision up to 3.75 megapixels, and introduces a completely new tokenizer. Pricing remains the same as Opus 4.6 at $5 and $25 per million tokens.

But the numbers only tell part of the story. Opus 4.7 is much better at handling complex and long-running tasks with consistency. It follows instructions more carefully and even finds ways to verify its own outputs before responding. In simple terms, it can catch its own mistakes before you do. That is a meaningful shift in how these models behave.

One of the most interesting additions is task budgets. This feature gives Claude a rough token limit for an entire workflow, including reasoning, tool usage, and the final output. The model keeps track of a running countdown and adjusts its approach so it can prioritise what matters and finish properly within the limit. For developers building long-running autonomous systems, this brings a level of control that has been missing.

Another update that did not get much attention is the tokenizer. The new tokenizer can use about 1x to 1.35x more tokens depending on the type of content. This means costs could increase slightly in some cases, which is something teams using Claude at scale should be aware of.

Anthropic has also added new cybersecurity-focused safeguards. Opus 4.7 includes systems that automatically detect and block high-risk or restricted cybersecurity requests. At the same time, security professionals can apply for access if they need it for legitimate use cases like penetration testing or vulnerability research.

Along with the model, Anthropic also introduced Claude Design. This is a new product that lets users work with Claude to create visual outputs such as designs, prototypes, slides, and one-pagers. The market reacted quickly, with companies like Figma and Adobe seeing their stocks drop, which shows how seriously this launch is being taken.

OpenAI Built an AI Specifically to Discover New Drugs

OpenAI launched GPT Rosalind this week, a frontier reasoning model built specifically for biology, drug discovery, and translational medicine. It is named after Rosalind Franklin, the scientist whose X ray crystallography work played a key role in uncovering the structure of DNA. The model is already being used by companies like Amgen, Moderna, and the Allen Institute.

It connects researchers to more than 50 scientific tools and data sources through a free Codex plugin, allowing it to run experiments, analyse molecular data, generate hypotheses, and iterate on drug candidates on its own. The speed at which it can operate is far beyond what any human research team could achieve. Unlike general purpose models that are adapted for science, GPT Rosalind has been built specifically for this domain and trained on biological and chemical datasets, clinical trial data, and decades of pharmaceutical research.

The bigger picture here is hard to ignore. Developing a single drug today usually takes 10 to 15 years and can cost over 2 billion dollars before it ever reaches patients. Most compounds fail along the way. Many promising ideas never make it from the lab to clinical trials, not because the science is wrong, but because the process of testing, iterating, and documenting is painfully slow.

GPT Rosalind is designed to address exactly these challenges. It can help identify which molecules are worth pursuing, predict how they might interact with biological systems, flag potential side effects early, and speed up the documentation and analysis that slows down every stage of development. This is not just a small step into healthcare. It is a serious push toward using AI to dramatically shorten the time it takes to bring new treatments to market.

If even part of this works as expected, the impact could be huge. Treatments that people might have expected to see in 2040 could arrive much sooner, potentially within the next decade.

Cursor Is Now Worth $50 Billion, It's 3 Years Old

Cursor, the AI coding startup built by Anysphere, is in advanced talks to raise around 2 billion dollars at a valuation of over 50 billion. Andreessen Horowitz is expected to co-lead the round, with Nvidia also planning to participate. Other firms like Thrive Capital and Battery Ventures are in discussions to join as well. The round is already oversubscribed, meaning demand from investors is higher than the available allocation. If it goes through, it would nearly double Cursor’s previous valuation of 29.3 billion, which was set just five months ago.

To put that into perspective, Salesforce took about 10 years to reach a 10 billion valuation. Cursor has crossed that level in less than 18 months.

The pace at which Cursor has grown is unlike anything seen before in enterprise software. The company reached 100 million in annualised revenue in January 2025, 500 million by June, 1 billion by November, and 2 billion by February 2026. Going from zero to 2 billion ARR in roughly three years makes it the fastest scaling B2B software company on record, ahead of companies like Slack, Zoom, and Snowflake.

It was founded by four MIT graduates in their mid twenties. Today, Cursor has more than one million paying customers, around two million total users, and about 50,000 enterprise teams. Nearly 70 per cent of the Fortune 1000 is represented in its customer base.

What makes this more than just another funding story is what it says about the direction of the software industry. More than half of the code on GitHub is now generated or assisted by AI. Around 90 per cent of developers use at least one AI tool at work. The biggest growth is happening in enterprises, where companies are moving from letting individual developers experiment with AI tools to rolling them out across entire engineering teams.

Cursor did not just build a better code editor. It showed up at the exact moment when writing software shifted from being a purely human task to a collaboration between humans and AI. It became one of the defining tools of that shift.

The company expects to end 2026 with an annualised revenue run rate of over 6 billion, which would mean tripling its revenue within a year. At a 50 billion valuation on 2 billion in current revenue, investors are clearly betting that this kind of growth will continue. Microsoft, with GitHub Copilot as a direct competitor, is watching closely.

At this point, the question is not whether AI coding tools will dominate software development. That is already happening. The real question is who will end up owning this market, and right now, a three-year-old startup built by four MIT graduates is leading the race.

Tesla Just Completed Its Most Important Chip Ever, And the Stock Jumped 8%

Tesla shares jumped nearly 8% this week, closing at $391.95, after Elon Musk shared that the company’s next-generation AI5 chip has reached a major engineering milestone and is now close to full production. This chip is not just another upgrade. It is expected to power everything Tesla is building next, from more advanced Full Self Driving to improved Optimus robots and a broader range of autonomous systems that other car companies are still far from delivering.

Tesla also announced plans for two new Terafab facilities in Austin, Texas. One will focus on vehicle and robot chips, while the other will be built for orbital data centres. This project is being developed in partnership with SpaceX, with Intel joining as a manufacturing partner. The idea of orbital data centres is especially interesting because it suggests computing infrastructure that operates above Earth’s atmosphere, without the usual constraints of land-based systems like power limits, regulations, and physical space.

What stands out most from this week is what it says about Tesla’s direction. This is no longer just an electric car company. It is becoming a fully integrated technology player that designs its own chips, builds its own factories, trains its own AI models, develops robots, and now even works on space-based computing infrastructure alongside a private space company.

Tesla also rolled out a spring software update, introducing a $99 per month subscription for Full Self Driving, along with voice-based integration of xAI’s Grok chatbot. Each layer of the system, from hardware to software to vehicles and beyond, is being built internally and connected together.

The stock movement reflects that shift in perception. Investors are not just valuing a car company anymore. They are starting to price in a much larger vision that looks like the foundation of a future Musk has been talking about for years. Whether or not that vision fully plays out, the progress is real, the infrastructure is being built, and even established players like Intel are now part of the bet.

Why This Week Matters

AI is no longer just about better models. It is about who controls entire industries.

OpenAI is moving into biology and drug discovery with a purpose-built model, which shows that AI is now targeting some of the most complex and high-value problems in the world. If this works even partially, it could compress decades of medical progress into years and change how new treatments are developed.

At the same time, Cursor’s rise shows how fast AI can reshape an existing market. Software development has already shifted from a human-only activity to a human and AI collaboration. The speed at which Cursor is growing suggests that the tools defining this new way of building software are still up for grabs, and the winners will be decided quickly.

Tesla’s updates point to something even bigger. Companies are no longer just building products. They are building entire stacks from chips to software to infrastructure. Tesla designing its own hardware, running its own AI, and now exploring space-based computing shows how far this level of control can go.

Put together, these stories show a clear pattern. AI is moving from being a layer on top of existing industries to becoming the core infrastructure those industries run on.

The companies that win will not just have the best models. They will control the full system around them, from data and compute to distribution and real world deployment.

At QuensultingAI, this is exactly what we help teams navigate. Whether it is building AI agents, automating workflows, or integrating AI into real business operations, the goal is not just to use AI, but to actually turn it into leverage.

The shift is already happening. The only question now is who ends up owning each layer.