Structure of the Case
The case begins by describing Jeff Hawkins’ life history, which provides interesting clues to the
making of a serial entrepreneur. Jeff Hawkins has two (professional) passions in his life: mobile com-
puting and neuroscience. Hawkins’ strong interest in brain theory was first aroused by a 1979 Scientific
American article written by Nobel Laureate Francis Crick (co-discoverer of DNA’s double-helix struc-
ture with Jim Watson). However, Hawkins’ early efforts to study how the brain works were rebuffed
by his supervisors at Intel, and his subsequent application to pursue a PhD in computer science was
rejected by MIT. Later, he enrolled in and then left the biophysics doctoral program at the University of
California, Berkeley after the faculty refused to sponsor his thesis proposal on a theoretical mathemati-
cal approach to understanding the brain.
Hawkins eventually found a more supportive environment at GRiD systems, where he created a
high-level programming language for application development, the handwriting recognition software
Grafitti, and the GRiDPAD (a tablet laptop computer) during his tenure with the company. In the early
1990s, Hawkins left GRiD to start his first new venture, Palm Computing (maker of the PalmPilot),
only to leave after Palm was acquired by 3Com in 1997. His second venture was Handspring, where
he launched the Visor and then the Treo (an early smartphone). When Handspring merged with
Palm’s hardware division in 2003, Hawkins decided it was finally time to return to his passion for
neuroscience, and he founded the nonprofit Redwood Neuroscience Institute. Listed under Additional
ments). Dileep George developed the mathematical breakthrough that enabled the operationalization
of Hawkins’ brain theory, and initially served as CTO (but is now on extended leave from his posi-
tion). Donna Dubinsky, Hawkins’ long-time business partner (Palm and Handspring), was the third
co-founder and the company’s first CEO. In May 2012, Numenta hired Rami Branitzky to take over
as CEO and oversee the company’s transition from the research and development stage into product
MHE-FTR-025
1259420477
REV: JANUARy 3, 2014
Grok: Action Intelligence for Fast Data1
TEACHING NOTE
Teaching Note —Grok: Action Intelligence for Fast Data
The case goes on to discuss the progress Numenta has made since its inception, including its cur-
rent business strategy and corporate partnerships. In 2007, Numenta released NuPIC, a platform for
HTM implementation that ran on Windows, Linux, and Mac OS X. Initially, Numenta provided a dual-
opment company to a commercial enterprise. Early adopters are exploring applications ranging from
video monitoring and image analysis, to optimizing energy reserve markets and monitoring wind farm
operations.
The final section of the case addresses the challenges Grok faces currently. Whether or not software
patents should be permitted has been a subject of great debate (the State Street Bank decision) and
port natural growth. Options that Hawkins might consider are selling out to a larger firm (acquisition),
holding an initial public offering, securing private equity funding, or forming strategic alliances to
procure much-needed resources.
Suggested Questions
ANAlySiS: FocuS oN ExtErNAl ANd/or iNtErNAl ENViroNmENtS
1. What is Hierarchical Temporal Memory (HTM)? How does it work? What are its potential / most
promising business applications?
2. What does the invention of HTM teach us about innovation?
3. What are Grok’s key resources and capabilities?
4. (Advanced) Can Grok’s resources and capabilities be a source of sustainable competitive
advantage? Why or why not?
FormulAtioN: FocuS oN BuSiNESS, corporAtE,
ANd/or GloBAl StrAtEGy
5. What business model should Grok implement? Explain.
6. Based on the industrial life cycle model, what should Grok’s strategic goals be at this stage of
development?
Teaching Note —Grok: Action Intelligence for Fast Data
implEmENtAtioN: FocuS oN rEcommENdAtioNS
ANd How to ExEcutE tHEm
7. What should Jeff Hawkins do to finance future growth for Grok?
8. How should Hawkins structure the Grok organization to best achieve its strategic goals?
ANAlySiS: FocuS oN ExtErNAl ANd/or iNtErNAl ENViroNmENtS
1. What is Hierarchical Temporal Memory (HTM)? How does it work? What are its potential / most
promising business applications?
Hierarchical Temporal Memory (HTM) is a machine-based learning model built on Hawkins’ mem-
ory prediction theory of how brains work, which was first laid out in his 2004 book on intelligence
(for a diagram of the human brain, see www.brainwaves.com/brain_diagram.html). HTM’s way of
addressing the key issue in building intelligent machines—knowledge representation—is fundamen-
tally different from how AI approaches it. AI’s way of tackling the problem is to build a huge database
of knowledge about the world. HTM, on the other hand, works by feeding the system with sequences
incapable of doing.
2. What does the invention of HTM teach us about innovation?
The case tells us what it takes, first of all, to create a great invention, and second to transform that
invention into a successful innovation. The role of the entrepreneur is highlighted in this case. Hawkins
has been chasing his “brain theory” dream all his life, displaying great passion and commitment. He
has pursued his dream in spite of being told by leading experts at Intel, MIT, and UC-Berkeley that
Teaching Note —Grok: Action Intelligence for Fast Data
In the market-technology 2×2 matrix, HTM is clearly a new technology rather than an existing one.
The market dimension is not as clear. HTM could be applied in existing markets such as real data
analysis; it could also be used to create some new market (for example, through the integration of
3. What are Grok’s key resources and capabilities?
Grok’s tangible resources include the financial backing of Jeff Hawkins and Donna Dubinsky, in
addition to a few other investors. Its geographical location is also a critical resource since Grok lies in
the heart of Silicon Valley, arguably the most important technology cluster in the United States (see
Case Exhibit 1 for a map of technology companies in Silicon Valley). Silicon Valley provides Grok with
access to a great pool of talented people, as well as a large amount of future venture capital.
Another important intangible resource is Grok’s HTM technology, which appears to be superior
to AI in solving complex problems. Grok can increase its intellectual property by continuing its com-
mercialization efforts and expanding the network of developers working with the HTM platform.
The company has increased its number of HTM-related patents to six in recent years, which further
strengthens its intellectual property space.
4. (Advanced) Can Grok’s resources and capabilities be a source of sustainable competitive
advantage? Why or why not?
Applying the VRIO framework (from Chapter 5), we can see that Grok’s tangible and intangible
resources are valuable. The financial backing provided by Grok’s founders and other investors enables
the business to grow. The company’s physical location provides access to both brain power and addi-
Teaching Note —Grok: Action Intelligence for Fast Data
tional investment capital. Grok’s human capital is perhaps the key to its success, as is the case in any
technology-intensive industry. Jeff Hawkins’ personal brand image helps Grok gain legitimacy and
support from important stakeholders. The joint managerial skills of Hawkins and Dubinsky assure
astute leadership. Meanwhile, the HTM technology and recent patents are the basis for Grok’s existence.
tive algorithms that have the same effect (for example, in reverse-engineering). This implies that Grok
should protect its intellectual property carefully, perhaps by building a “patent fence.” Overall, Grok
possesses hard-to-imitate resources that can be used as the basis for building a sustainable strategic
competitive advantage.
FormulAtioN: FocuS oN BuSiNESS,
corporAtE, ANd GloBAl StrAtEGiES
5. What business model should Numenta implement? Explain.
Grok’s future depends on how it decides to leverage its resources and capabilities into a profit-
able business model. The company’s initial business model—a proprietary platform and dual licens-
ing structure—facilitated the rapid development of the NuPIC software by building a community of
open-source developers who were granted access under a free research license. To support developers
in their efforts, Grok’s website featured multiple tools such as a forum, wikis, and a blog written by
Grok released its second-generation software, based on newer HTM algorithms, for beta testing in
2011. In anticipation of this commercial release, the company ceased active maintenance of NuPIC and
its related vision applications. Existing NuPIC users had the choice to either adopt the newer version
(at the commercial rate), or continue using the legacy software without the benefit of Grok’s support.
Teaching Note —Grok: Action Intelligence for Fast Data
One business model that Hawkins is likely to consider is what Microsoft did with MS-DOS: licens-
ing it non-exclusively for a small fee to all comers. This enabled Microsoft to make MS-DOS the stan-
dard operating system for personal computing. In this vein, Grok has already entered a number of
6. Based on the industrial life cycle model, what should Grok’s strategic goals be at this stage of
development?
Grok’s HTM software is in the introduction stage, where firms invest in R&D to create a new product
and form an initial customer base (early adopters). Only a few innovators are active in the market, and
products are positioned based on features and performance as opposed to affordability. Prices tend to
be high to cover the expenses associated with R&D, small-scale production, and initial product launch.
implEmENtAtioN: FocuS oN rEcommENdAtioNS
ANd How to implEmENt tHEm
7. What should Jeff Hawkins do to finance future growth for Numenta and its HTM technology?
Although both Hawkins and Dubinsky are independently wealthy, they cannot fund Grok’s activi-
ties in perpetuity. To grow, Grok needs to create a reliable cash flow. Likely options include an IPO or
an acquisition by a larger firm (such as IBM). In both cases, Hawkins would likely lose direct control
over the technology and the firm he founded.
Strategic acquisition can be attractive to entrepreneurs who seek immediate liquidity. In the case of
an IPO, CEOs and other early investors may not be able to sell immediately, and much can happen—
both good and bad—before the lock-up period on their shares expires. Selling the company enables
them to realize its value and exit the business more quickly. Another problem with IPOs is the cyclical
Teaching Note —Grok: Action Intelligence for Fast Data
7
Copyright © 2017 McGraw-Hill Education. All rights reserved. No reproduction, distribution, or posting online without the prior written consent
of McGraw-Hill Education.
truly learn from past experience as described in Hawkins’ memory prediction framework, we would
expect to see Hawkins avoid a replay of this potential scenario. In the second instance, the entrepreneur
immediately receives cash or unrestricted stock for the full value of the ownership interest, but does
not have a continuing role or direct financial interest in his company’s future growth. However, Grok
is Hawkins’ “baby,” the culmination of his lifetime pursuit of a theory of how brains work, and he is
unlikely to want to leave. Hawkins may also choose an IPO in consideration of Grok’s employees, since
strategic buyers often cut costs by eliminating duplicate departments and functions.
Now that the stock market is slowly picking up, especially in the technology sector, a few years from
now may be a good time for Grok to go public. With the independent wealth of its initial investors
8. How should Hawkins structure the Grok organization to best achieve its strategic goals?
With the shift toward commercialization, Grok has started to transition from a simple structure
toward a functional structure. In firms with a simple structure, the founders tend to make all the impor-
tant strategic decisions and run the day-to-day operations. The organization is relatively flat in nature,
with a low degree of formalization and specialization.
Engineering and Craig Vaughan as VP of Marketing and Product Management. Another sign of transi-
tion is that the company now has a five-member board of directors, including Hawkins and Dubinsky
and three outsiders.
Recent Updates
Grok underwent a leadership change subsequent to the writing of this case. Rami Branitzky is no
longer with the company, and Donna Dubinsky has resumed her role as CEO while remaining board
chair. In addition, Grok has added two new members to its senior management team: Subutai Ahmad
as VP of Engineering and Craig Vaughan as VP of Marketing and Product Management.
As part of its commercialization process, Grok has also formed a board of directors consisting of
Donna Dubinsky, Jeff Hawkins, Ed Colligan, Mike Farmwald, and Harry Saal. Brief biographies of all
senior managers and board members are provided in Exhibit TN-1.
Teaching Note —Grok: Action Intelligence for Fast Data
Additional Resources
1. For additional reading, see Jeff Hawkins’ book with Sandra Blakeslee (2004), on intelligence: How a
New understanding of the Brain will lead to the creation of truly intelligent machines (New york: Owl
Books).
2. http://ecorner.stanford.edu/authorMaterialInfo.html?mid=2217 (~57 min)
Stanford’s University’s Entrepreneurship Corner, Entrepreneurial Thought Leader Lecture: inside
the mind of a reluctant Entrepreneur. “Numenta’s Jeff Hawkins, a frequent company founder,
inventor, and product designer for Palm and Handspring, highlights lessons learned during his
tenure in technology. He also confesses that these accomplishments were mere way stations in his
30-year passionate pursuit of neuroscience.” Specific video segments are listed in the following for
your convenience.
the twenty-year plan. (3:16)
the History of palm and the part-time Entrepreneur. (06:07)
An Entrepreneur’s take on the Non-profit. (01:57)
Effective choices, Not longer days. (02:01)
Forgo the NdA: Find Support where you can. (02:38)
case Studies of Failure. (05:28)
palm’s Approach to Elegant design. (06:16)
3. http://www.ted.com/talks/lang/eng/jeff_hawkins_on_how_brain_science_will_change_
computing.html (20:16 min). TED talk: Jeff Hawkins on How Brain Science will change computing.
“Treo creator Jeff Hawkins urges us to take a new look at the brain—to see it not as a fast processor,
but as a memory system that stores and plays back experiences to help us predict, intelligently,
what will happen next.”
4. http://www.youtube.com/watch?v=G6CVj5IQkzk (22:09 min). Brain Science is about to
Fundamentally change computing. A brief talk on how Hawkins’ brain theory works.
5. Jeff Hawkins on Artificial intelligence. (60 min total). This video series provides a more detailed
discussion of how HTM works. Jeff Hawkins also touches on the potential applications of HTM
and a little bit on Numenta/Grok.
Part 1/5: http://www.youtube.com/watch?v=oozFn2d45tg. (09:47).
Part 2/5: http://www.youtube.com/watch?v=pjpWISEIl44. (09:57).
Part 3/5: http://www.youtube.com/watch?v=pngRexsrmGk. (09:02).
Teaching Note —Grok: Action Intelligence for Fast Data
Part 4/5: http://www.youtube.com/watch?v=GpekuvG4JMg. (06:38).
Part 5/5: http://www.youtube.com/watch?v=FthvUpHzwfA. (05:49).
Teaching Note —Grok: Action Intelligence for Fast Data
ExHIBIT TN-1 Grok’s Top Management Team
(continued)
continued to serve on the board until early 2009. Previously,
Donna spent 10 years at Apple Computer in a variety of sales,
the Harvard Business School. She currently serves on the
board of the Peninsula Open Space Trust, the Computer
Jeff Hawkins is an engineer, serial entrepreneur, scientist,
inventor, and author. He was a founder of two mobile
computing companies, Palm and Handspring, and was the
In 2002, he founded the Redwood Neuroscience Institute,
a scientific institute focused on understanding how the
In 2005, he co-founded Grok (formerly known as Numenta), a
startup company building a technology based on neocortical
theory. It is his hope that Grok will play a catalytic role in the
Cornell University in 1979. He was elected to the National
Academy of Engineering in 2003.
Engineering team. Prior to Grok, Subutai served as VP
Engineering at YesVideo, Inc. From 1999 until 2005, he
video analysis systems have been deployed internationally
the University of Illinois at Urbana-Champaign.
Most recently, Craig held global leadership roles in marketing
and sales at SAP—managing strategic marketing and sales
also held roles in strategy consulting and venture capital,
focusing on high-growth emerging companies.
in Marketing and Finance).
building the Radius brand to be synonymous with graphics
and graphics imaging at the outset of the desktop publishing
brand. Ed left Palm to found Handspring, where he and his
partners created the forbearer of all future smartphones: the
Handspring Treo. Ed drove the transaction that reunited Palm
Ed now spends his time investing in and mentoring
entrepreneurs with the goal of helping establish the next
Teaching Note —Grok: Action Intelligence for Fast Data
ExHIBIT TN-2 A VRIO Analysis of Numenta
Numenta’s Resources and Capabilities
… are they?
Valuable
V
Rare
R
Costly to
Imitate
I
Organized to
capture value
O
Source: http://www.numenta.com/about.html.
ExHIBIT TN-1 Grok’s Top Management Team (Continued)
of media processors that was acquired by ATI Technologies,
Epigram—a creator of advanced semiconductor home
Harry Saal, Director
Chairman of ZERO1, the Arts & Technology Network
In 2002, Dr. Harry J. Saal was chosen by the U.S. Department
of Justice to lead the Technical Committee charged with
1993 through 1995, Harry served as founding CEO and
President of Smart Valley, Inc., a non-profit organization
Harry holds a BA, MA, and PhD in Physics from Columbia
University. Harry serves on the board of the American
Teaching Note —Grok: Action Intelligence for Fast Data
Endnotes
1. https://www.groksolutions.com/index.html.