TechCast - A Virtual Think Tank Tracking the Technology Revolution
Home Forecasts Library Services FAQ About TechCast   Search

Breakthrough Analysis

Utility Computing

by William E. Halal
Monday, October 15, 2007


Event | Data Points | Forecast Data Analysis | Trend Analysis | Articles  

After a false start earlier, “utility computing,” or “computing on demand,” is being taken seriously now as IBM enters the field. If the challenge of managing this complex process can be overcome, computing power could become as convenient and affordable as electricity, water, and phone service. “We think this is the third major computing revolution,” said a Forester analyst. (BusinessWeek, 8/25/03) TechCast data show arrival time at 2010.

Event Being Forecast

30% of computer use is obtained from companies that provide services (computational power, software, security, etc.) on demand.


Selected Data Points to Consider

IDC estimates it will take until 2007 to make the technology available, and IBM thinks it will be ready about 2007-09. (BusinessWeek, 11/1102)

• Market demand is expected to grow from $8.6 billion in ’03 to $25 billion in ’07, with penetration rising from 15% today to 30% in ’07. ( 5/28/04)

  Back to Top of Utility Computing

Forecast Data Analysis

    Mean Std Dev N (# Experts)
Most Likely Year 2011 3 44
Market Size ($B) 270 246
Confidence (%) 66 14

In addition, 7 experts predicted that this event would never occur; mean confidence: 53%; std. dev.: 22.2.

Note: "Most Likely Year" is the year when this technology is expected to reach the adoption level stated under "Event Being Forecast" in industrial nations like the US, EU, and Japan. "Confidence" is the confidence our experts place in this forecst. "Market Size" estimates the potential market demand when this rechnology matures; global figures are not available, so this estimate is for the U.S. economy.

Forecast Chart

Back to Top of Utility Computing

Trend Analysis

PROS: Trends Driving this Event

  1. VALUE Utility computing provides access to processing power and software without buying and maintaining complex systems, it puts the responsibility for security on the provider, and can be accessed with a dumb terminal anywhere. It is also efficient because most servers only use 20% of their capacity and PCs use about 5%. Less than 10% of the capability of software is used by typical people. Fluor Corp. saved $85 million in IT costs using utility computing and was able to set up a new data center in 70 hrs rather than the normal 3 months. (, 5/28/04)

  2. PROVIDERS ARE APPEARING • SimDesk provides basic services (word processing, email, etc.)to millions of people in Houston, Chicago, and Indiana, and may expand into other states and entire nations. (Businessweek 7/5/04)
    • IBM is investing $ 10 B to make utility computing the heart of its corporate strategy, “E-Business on Demand,” and it is introducing “Workplace” as a server-based alternative to Microsoft Office. “This is exactly what customers want,” said the CEO, Samuel Palmisano. “It’s going to drive phenomenal productivity.” (Businessweek 11/11/02)
    • Hewlett-Packard and Sun are developing an “intraorganizational” approach in which corporate "computing networks" serve units within the organization. (BusinessWeek, 8/25/03)
    • uses only 10% of its computer capacity, so CEO Jeff Bezos is selling utility computing to customers like Microsoft. (BusinessWeek 11/13/06)
    • Sun Microsystems has developed a "Blackbox" that packs servers in a shipping container to offer 7 terabytes of computer power and 2 petabytes of data storage at 1 % of the cost of most servers. Sun expects the skyrocketing demand for online services used by desktops, laptops, smartphones,iPods, etc. to create a huge need for utility computing that can be set up easily at low cost. (Scientific American 8/07)

  3. COMPARABLE TO ELECTRIC UTILITIES The growth of utility computing parallels that of electric power. It took 2 decades for central power stations to emerge, which is roughly comparable to the emergence of utility computing. (, 11/9/04)

CONS: Obstacles Opposing this Event

  1. TECHNICAL LIMITATIONS Response time can be slow as commands have to travel between the user and the utility’s servers and back again.

  2. COMPLEXITY Managing central computing power to serve a variety of different client needs is daunting and requires standards and compatible systems. (BusinessWeek 8/25/03)

  3. LOSS OF CONTROL Individual buyers and corporate clients are often reluctant to become dependent on outside sources for something as strategic as computer services.

Back to Top of Utility Computing


  1. Can Computers Think?

  2. The Intelligent Internet

  Back to Top of Utility Computing

Copyright © 2005 TechCast, LLC All Rights Reserved | Terms of Use | Privacy Policy | Contact Us | Manage Your Account