Why Microsoft stock could see upto 20% more upside despite gaining 30% since earnings

Investor concerns over software-sector weakness and excessive artificial intelligence spending are easing, giving Microsoft shares fresh momentum as Wall Street gains confidence in the company’s ability to turn its AI investments into returns.

Bank of America analyst Tal Liani believes the rally has further room to run.

In a Tuesday note, Liani maintained his Buy rating on Microsoft and lifted his price target to $600 from $500, implying roughly 20% upside from current levels.

The new target is based on 28 times Bank of America’s calendar 2027 earnings-per-share estimate, compared with 24 times previously.

The adjustment reflects accelerating cloud growth and improving visibility into AI investment returns.

Microsoft shares have gained nearly 30% since the end of July and recently recorded their strongest six-day stretch since October 2025.

Azure and Copilot provide the evidence

The latest rally follows a strong fiscal fourth-quarter earnings report.

Azure revenue growth accelerated 43% year over year, while Microsoft's guidance for 45% growth in the current quarter exceeded Wall Street expectations.

The company is also seeing increasing adoption of its AI-powered Copilot products.

Liani noted that paid Copilot seats have surpassed 30 million, with net new additions more than doubling from the previous quarter.

For Liani, however, Microsoft's advantage goes beyond any individual AI model.

The company is increasingly building what he describes as a full-stack AI portfolio, combining its own models with offerings from third-party providers.

"Microsoft is building a broad portfolio of internal and external models, allowing customers to use the most cost-effective model for each task, while governing users actions," Liani said.

"Not every workload requires a complex and expensive frontier model, and Microsoft's approach helps optimize performance while reducing token consumption."

A model-agnostic approach could improve AI economics

Rather than relying on a single model provider, Microsoft can route workloads according to their complexity and cost.

By employing a mix of internal and external models, "Microsoft can reserve the largest and most expensive models for complex tasks while serving high-volume, product-specific workloads more efficiently," Liani wrote.

That approach could help Microsoft reduce its exposure to the economics of any one provider while allowing Copilot to remain flexible as AI models evolve.

Liani also pointed to Microsoft's MAI-Code-1-Flash model, which he said delivers performance comparable to GPT-5.6 for common Excel tasks at a lower cost.

The strategy means Copilot's value does not "depend exclusively" on Anthropic, OpenAI or another individual model provider.

Hardware efficiency adds another layer

Microsoft is also attempting to improve AI economics at the infrastructure level.

Liani said engineering improvements across its CPU and GPU fleet, combined with software optimization, have increased throughput for Copilot workloads fourfold since the beginning of the year.

"Engineering improvements across Microsoft's CPU and GPU fleet, combined with software optimization, increased throughput for Copilot workloads by 4x since the beginning of the year," Liani wrote.

Microsoft is also developing custom silicon to reduce inference costs.

Its Maia 200 chip, which powers Microsoft AI models, is up to 40% cheaper to operate than traditional Nvidia hardware, according to the analyst.

These efficiency gains could become increasingly important as Microsoft brings additional AI capacity online.

Liani now expects Azure to grow 41.8% in fiscal 2027, compared with 39.9% in fiscal 2026.

"Continued execution on the capacity buildout, faster deployment and greater efficiency increase our confidence in Microsoft's ability to sustain Azure growth," he said.

The key question for investors is no longer simply whether Microsoft can spend enough to compete in AI.

It is whether the company can keep converting that spending into faster cloud growth and stronger economics. For now, Bank of America believes the answer is increasingly yes.